cfpark00/toy-multistep-nn_20-na_5-nab_30-seed_2

这是一个包含文本提示和完成文本的数据集，同时还包括了每个示例中遮蔽的单词数量。数据集分为训练集、测试集(test_rl)和测试集(test)，每个集合都包含262144个示例。数据集适用于文本生成任务，特别是那些需要根据提示生成文本的任务。

This dataset includes text prompts and their corresponding completions, as well as the number of masked words in each example. The dataset is split into three parts: a training set, a test set (test_rl), and another test set (test), each containing 262144 examples. The dataset is suitable for text generation tasks, particularly those that involve generating text based on prompts.